Neocortical neurons in vivo are spontaneously active and intracellular
recordings have revealed strongly fluctuating membrane potentials ari
sing from the irregular arrival of excitatory and inhibitory synaptic
potentials. In addition to these rapid fluctuations, more slowly varyi
ng influences from diffuse activation of neuromodulatory systems alter
the excitability of cortical neurons by modulating a variety of potas
sium conductances, In particular, acetylcholine, which affects learnin
g and memory, reduces the slow afterhyperpolarization, which contribut
es to spike frequency adaptation. We used whole-cell patch-clamp recor
dings of pyramidal neurons in neocortical slices and computational sim
ulations to show, first that when fluctuating inputs were added to a c
onstant current pulse, spike frequency adaptation was reduced as the a
mplitude of the fluctuations was increased. High-frequency, high-ampli
tude fluctuating inputs that resembled in vivo conditions exhibited on
ly weak spike frequency adaptation. Second, bath application of carbac
hol, a cholinergic agonist, significantly increased the firing rate in
response to a fluctuating input hut minimally displaced the spike tim
es by < 3 ms, comparable to the spike jitter observed when a visual st
imulus is repeated under in vivo conditions. These results suggest tha
t cholinergic modulation may preserve information encoded in precise s
pike timing, but not in interspike intervals, and that cholinergic mec
hanisms other than those involving adaptation may contribute significa
ntly to cholinergic modulation of learning and memory.